AnsweredAssumed Answered

AD8067 Spice model vs data sheet

Question asked by jcx on Dec 12, 2012
Latest reply on Jan 8, 2013 by jcx

Playing with Audio DAC I/V op amp selection I tried Analog's AD8067 Spice model and found some problems when compared to the data sheet parameters, plots

 

the model's ~74 dB Avol doesn't come anywhere near the data sheet DC open loop gain of 100 dB min, 119dB typical

 

the .cir file comment shows the problem

  * GAIN STAGE & POLE AT 17 kHz

 

the data sheet gain vs frequency plot doesn't show any such high frequency pole limiting DC gain, the table specs tell a different story too

 

one graph suggests +5 V single supply, 30 mA load could drag the gain down that far – but is that really the performance you want to build in to the default Spice model?

 

 

The output Z agrees well with the data sheet – everywhere it hardly matters, but the Spice model has a 60 MHz resonance with >2x the data sheet |Z|, drops to a flat R above 100 MHz – while the data sheet output Z plot keeps rising

 

its sorta useful to capture the Zout performance near the gain intercept frequency to see C load stability, peaking in the sim – even better to see the model be close for the next decade beyond

 

 

the AD8067.cir Vos of 1.5 mV is also annoying – according to the data sheet I can expect the factory to replace any AD8067 with such large Vos –  greater than the data sheet max Vos, way larger than the ~200-300 uV typical as verified by the Vos histogram in the data sheet

unless you're piling on worst case assumptions - say Vos tempco and high operating T  I don't see how such large Vos
belongs in the default Spice model

 

 

and I also like to point out that Ccm modeling parts do not connect to gnd in a real op amp – they are resolved with respect to the supply pins – how hard is it to make this better?

or even add reverse biased diodes with C voltage dependency from inputs to supply pins?

Outcomes